Abstract
Bacterial meningitis occurs in approximately 0.4 neonates per 1000 live births. Virtually all organisms that cause neonatal infection or sepsis can result in central nervous disease with severe consequences to the developing brain. Early appropriate therapy is mandatory to improve both short- and long-term outcomes. This is possible only by the timely recognition of its occurrence, thus making performance of a lumbar puncture for cerebrospinal fluid analysis and culture the key to rapid institution of effective antimicrobial therapy.
Keywords
antibiotic, meningitis, neonate
- •
As many as 40% of infants with meningitis who have a gestational age of > 34 weeks do not have a positive blood culture at the time of their diagnosis.
- •
Urine culture is an important part of a sepsis evaluation since urinary tract infection is relatively common in neonates who are greater than 72 hours of age.
- •
For late-onset sepsis, a penicillinase-resistant, semisynthetic penicillin such as oxacillin or nafcillin in combination with an aminoglycoside is the preferred choice.
- •
The treatment of gram-negative meningitis initially includes the addition of a third or fourth generation cephalosporin such as cefotaxime or cefepime 77,78 or a carbapenem antibiotic such as meropenem.
- •
For meningitis due to gram-negative bacilli, the duration of antimicrobial therapy is a minimum of 21 days.
- •
For meningitis due to group B streptococcus, a minimum of 10 days of antimicrobial therapy is recommended.
- •
Among infants with gram-negative enteric meningitis approximately 20% to 30% of affected infants die, and neurologic sequelae are found in 35% to 50% of survivors.
Bacterial meningitis occurs in approximately 0.4 neonates per 1000 live births. It is defined as inflammation of the meninges that is manifested by an elevated number of white blood cells in the cerebrospinal fluid (CSF). It often is associated with elevated protein content and a low glucose concentration in CSF. Meningitis generally results as a consequence of hematogenous dissemination of bacteria via the choroid plexus and into the central nervous system (CNS) during a sepsis episode. Invasion of the meninges occurs in about 10% to 20% of infants with bacteremia. Rarely, meningitis develops secondary to extension from infected skin through the soft tissues and skull as may occur with an infected cephalohematoma or direct spread from skin surfaces, as in infants with myelomeningoceles or other congenital malformations of the neural tube. In addition, ventriculoperitoneal shunts or ventricular reservoirs may be the primary site of infection. A potential but infrequent complication of meningitis is brain abscess that results from hematogenous spread of bacteria into tissue that has incurred anoxic injury or severe vasculitis with hemorrhage or infarction.
Virtually all organisms that cause neonatal infection or sepsis can result in central nervous disease with severe consequences for the developing brain. A list of these pathogens is provided in Box 11.1 . It is imperative that a correct and timely diagnosis with a specific organism be made because treatment decisions vary by causative agent.
|
|
|
|
|
|
|
|
For others, see Giacoia GP. Uncommon pathogens in newborn infants. J Perinatol. 1994;14:134–144.
For a more complete listing, see Palazzi DL, Klein JO, Baker CJ. Bacterial sepsis and meningitis. In: Remington JS, Klein JO, Wilson CB, Baker CJ, eds. Infectious Diseases of the Fetus and Newborn Infant. 6th ed. Philadelphia: WB Saunders; 2006:247–295.
The case of a preterm infant is presented and discussed to illustrate and highlight the multifaceted nature of this disease. The objective of this chapter is to review the current management of neonatal bacterial meningitis, in the hope of ameliorating the destructive nature of many of these organisms and ultimately improving the outcome of these high-risk infants.
A preterm infant weighing 1004 g was born at 28 weeks’ gestation to a 24-year-old mother by cesarean section. The pregnancy was complicated by premature rupture of membranes 2 weeks before delivery, and the mother developed intrapartum fever and was diagnosed with chorioamnionitis. She received antenatal steroids and antimicrobial therapy consisting of ampicillin and gentamicin. At delivery, the infant was floppy with poor respiratory effort, and he required intubation and admission to the neonatal intensive care unit (NICU). Apgar scores were 3 at 1 minute and 7 at 5 minutes. The infant’s vital signs were normal, and antimicrobial therapy with ampicillin and gentamicin was initiated after a blood culture was obtained. Hyaline membrane disease was diagnosed and the infant received exogenous surfactant therapy.
Question 1: What Risk Factors Predispose This Infant to Have Early-Onset Bacterial Meningitis?
Because meningitis is a complication of bacteremia; the risk factors are similar to those that contribute to neonatal sepsis—namely prematurity, prolonged rupture of fetal membranes (≥18 hours), maternal urinary tract infection, and maternal intrapartum fever or chorioamnionitis. Immune dysfunction as well as lack of transplacentally acquired maternal immunoglobulin G (IgG) antibodies in premature infants also may increase risk of sepsis and CNS infection. Recently, lower neonatal 25-hydroxyvitamin D levels have been associated with early-onset sepsis.
Likewise, clinical signs suggestive of bacterial meningitis are similar to those of neonatal sepsis. In the full-term infant, fever, lethargy, hypotonia, irritability, apnea, poor feeding, high-pitched cry, emesis, seizures, and bulging fontanelle are prominent clinical signs, whereas in preterm infants, respiratory decompensation consisting of an increased number of apneic episodes predominates. Neonates with meningitis are never “asymptomatic.”
The widespread and routine use of intrapartum antimicrobial chemoprophylaxis since 1996 has significantly reduced the rate of early-onset group B streptococcal (GBS) infection by more than 70%. Fortunately there has not been a reciprocal increase in early-onset bacterial infections caused by gram-negative organisms among all newborns in the United States. However, among very low-birth-weight (VLBW) infants with birth weight of 1500 g of less, a shift toward more gram-negative infections has occurred. Among the NICUs of the National Institute of Child Health and Human Development (NICHD) Neonatal Research Network centers, intrapartum antimicrobial chemoprophylaxis resulted in a significant decrease in early-onset GBS infection while the rate of infections caused by Escherichia coli increased significantly from 3 to 7 cases per 1000 live births. More recently, another NICHD Neonatal Research Network study documented rates of culture-confirmed early-onset sepsis among almost 400,000 live births at Network centers. The overall rate of early-onset sepsis, defined as a positive blood or cerebrospinal fluid bacterial culture at less than 72 hours of age, was 0.98 infections per 1000 live births with rates inversely related to birth weight (BW; 401–1500 g BW, 10.96/1000; 1501–2500 g BW, 1.38/1000; >2500 g BW, 0.57/1000). Among cases of early-onset meningitis, 44% were due to E. coli while only 19% were due to GBS. The majority of E. coli isolates were resistant to ampicillin, an antibiotic that is often used for intrapartum GBS chemoprophylaxis. Even more concerning was the finding that 3% of E. coli isolates were resistant to the third-generation cephalosporin agents.
Question 2: Do Infants With Meningitis Have Positive Blood Cultures?
As many as 40% of infants with meningitis who have a gestational age of 34 weeks or more do not have a positive blood culture at the time of their diagnosis. Similarly, among VLBW infants, almost one-half of cases of meningitis occur with sterile blood cultures. Therefore it is imperative that a lumbar puncture be performed if sepsis or meningitis is suspected. Evaluation of CSF indices and Gram stain not only will establish a diagnosis but also will help guide initial therapy. Normal CSF indices are provided in Table 11.1 .
| Birth Weight (g) | Age (Days) | No. of Patients | Red Blood Cells (/mm 3 ) Mean ± SD (Range) | White Blood Cells (/mm 3 ) Mean ± SD (Range) | Polymorphonuclear Leukocytes (%) Mean ± SD (Range) | Glucose (mg/dL) Mean ± SD (Range) | Protein (mg/dL) Mean ± SD (Range) |
|---|---|---|---|---|---|---|---|
| Preterm Neonate a | |||||||
| ≤1000 | 0–7 | 6 | 335 ± 709 (0–1780) | 3 ± 3 (1–8) | 11 ± 20 (0–50) | 70 ± 17 (41–89) | 162 ± 37 (115–222) |
| 1001–1500 | 8–28 | 17 | 1465 ± 4062 (0–19,050) | 4 ± 4 (0–14) | 8 ± 17 (0–66) | 68 ± 48 (41–89) | 159 ± 77 (95–370) |
| 29–84 | 15 | 808 ± 1843 (0–6850) | 4 ± 3 (0–11) | 2 ± 9 (0–36) | 49 ± 22 (41–89) | 137 ± 61 (76–260) | |
| 0–7 | 8 | 407 ± 853 (0–2450) | 4 ± 4 (1–10) | 4 ± 10 (0–28) | 74 ± 19 (41–89) | 136 ± 35 (85–176) | |
| 8–28 | 14 | 1101 ± 2643 (0–9750) | 7 ± 11 (0–44) | 10 ± 19 (0–60) | 59 ± 23 (41–89) | 137 ± 46 (54–227) | |
| 29-84 | 11 | 661 ± 1198 (0–3800) | 8 ± 8 (0–23) | 11 ± 19 (0–48) | 47 ± 13 (41–89) | 122 ± 47 (45–187) | |
| Full-Term Neonate b | |||||||
| 0–30 | 108 | ≤1000/mm 3 | 7.3 ± 13.9 (0–130) median 4 | 0.8 ± 6.2 (0–65) median 0 | 51.2 ± 12.9 (62% of serum glucose) | 64.2 ± 24.2 | |
a From Rodriguez AF, Kaplan SL, Mason EO Jr. Cerebrospinal fluid values in the very low birth weight infant. J Pediatr . 1990;116:971–874.
b From Ahmed A, Hickey SM, Ehrett S, et al. Cerebrospinal fluid values in the term neonate. Pediatr Infect Dis J . 1996;15:298–303.
Meningitis in preterm infants admitted to the NICU with respiratory distress syndrome is very uncommon. Therefore performance of a lumbar puncture in these infants in whom sepsis is not suspected is not mandatory. Similar data are available for full-term infants. However, if the blood culture yields a pathogenic organism, then evaluation of CSF should be done. A lumbar puncture is contraindicated when there is cardiorespiratory instability. Delay in performing a lumbar puncture only delays a potential diagnosis of meningitis and can lead to prolonged and possibly inappropriate antibiotic use.
The infant was extubated and continuous positive airway pressure therapy was started on the first day of age. Trophic feedings were initiated on the second day of age, and a percutaneous intravenous central venous catheter was placed for parenteral nutrition. She achieved full enteral feedings on the 20th day. Over the subsequent 2 days, she developed lethargy, hyperglycemia, and increased episodes of apnea that resulted in reinitiation of mechanical ventilation. Two blood cultures were obtained, and antimicrobial therapy with nafcillin and gentamicin was initiated.
Question 3: What Is the Optimal Evaluation for Possible Late-Onset Sepsis in Preterm Infants in the NICU?
Infants suspected of having late-onset sepsis in the NICU should have a complete evaluation that consists of a complete blood cell (CBC) count, a urinanalysis, and culture. In critically ill hospitalized neonates, it is very difficult to distinguish infection from noninfectious clinical deteriorations; however, there should be a low threshold for performing a lumbar puncture. Unfortunately, there is no laboratory or clinical finding that has a sensitivity of 100% for the diagnosis of neonatal sepsis. Such laboratory tools as CBC count; C-reactive protein; interleukin (IL)-6, IL-8, IL-10; and procalcitonin have suboptimal sensitivity and specificity to replace the blood culture as the gold standard, but these tests may be useful to support a diagnosis of infection when their results are abnormal and accompanied by clinical signs of infection. Polymerase chain reaction for detection of bacterial and fungal DNA ultimately may lead to an earlier diagnosis.
It is important to obtain a CBC count with platelets for reasons other than diagnosis. Neonatal sepsis may result in neutropenia, which is associated with a high mortality rate. The finding of an absolute neutrophil count of 500/mm 3 or less may prompt the administration of intravenous immunoglobulin (IvIg 750 mg/kg); the use of IvIg has been associated with improvement in the peripheral neutrophil count presumably from improved neutrophil egress from the bone marrow in infants with sepsis. Routine use of IvIg infusions for suspected sepsis is not recommended, however, as studies have not demonstrated benefit in early or late morbidity or mortality. Recombinant granulocyte or granulocyte-macrophage colony-stimulating factors can be considered if IvIg is unsuccessful in improving the neutrophil count. Another reason for performance of a CBC count is evaluation of the platelet count because disseminated intravascular coagulation may result in severe thrombocytopenia. In addition, thrombocytopenia may be an early marker of disseminated candidiasis.
Debate continues as to whether multiple blood cultures should be performed. Certainly with bacterial organisms that are frequent blood culture contaminants, such as coagulase-negative staphylococci (CoNS), the diagnosis of sepsis is best confirmed by the finding of two or more positive cultures from multiple sites or body fluids that are normally sterile. Two peripheral vein blood cultures are recommended in an infant with a suspected late-onset infection. However, given the difficulty of obtaining blood cultures, this often means obtaining a blood culture from a central catheter and another from a peripheral blood vessel. The value of obtaining a blood culture from a central venous catheter has been debated, because a positive central line culture may represent colonization rather than true infection. The isolation of CoNS from only one blood culture when only one is obtained is problematic and of uncertain significance. Because many of these positive cultures represent contamination with skin microflora, the practice of obtaining only one blood culture often leads to prolonged and unnecessary antibiotic therapy. In addition, performance of two blood cultures may increase the likelihood of isolating a causative agent. This practice leads to more prudent antibiotic use—a major goal in the NICU where antimicrobial resistance is an emerging but preventable problem.
Urine culture is an important part of the evaluation because urinary tract infection is relatively common in neonates older than 72 hours of age. Urine should be obtained by suprapubic bladder aspiration whenever possible, and the finding of any growth is significant. Alternatively, a catheterized urine specimen may be obtained, recognizing that urethral or perineal bacterial or fungal contamination in these small infants may complicate the assessment of results. In general, the presence of at least 50,000 colonies per milliliter of a single organism is considered a true urinary tract infection, whereas lower colony counts are more indicative of contamination. In addition, microscopic analysis for evidence of pyuria is useful to support its diagnosis. Bag specimens should never be obtained for the evaluation of possible urinary tract infection.
A chest radiograph should be obtained if respiratory decompensation is present. A lumbar puncture should generally be performed in infants evaluated for possible late-onset sepsis for reasons stated in answer to Question 2. Risk factors for meningitis in preterm infants include low gestational age and prior bloodstream infection. In VLBW infants, the average age of late-onset meningitis is 26 days (median 19 days; range 4–102 days). Therapeutic decisions with regard to antibiotic choices can be made only if one knows whether the CNS is involved.
Question 4: What Is the Empirical Antimicrobial Choice for Possible Late-Onset Sepsis in the NICU?
In general, antimicrobial therapy for neonatal sepsis is dependent on the agents commonly seen in that particular nursery and their susceptibility pattern. For early-onset sepsis, ampicillin combined with an aminoglycoside, usually gentamicin, has been the empiric therapy of choice since group B Streptococcus , other streptococcal species, Listeria monocytogenes , and gram-negative bacilli predominate.
For late-onset sepsis, a penicillinase-resistant, semisynthetic penicillin such as oxacillin or nafcillin in combination with an aminoglycoside is the preferred choice. For CNS infections, nafcillin is preferred because of improved penetration. Because approximately 50% of all bloodstream infections are due to CoNS, some experts recommend vancomycin instead of a semisynthetic penicillin because CoNS are almost uniformly resistant to these agents. This practice has led to widespread use of vancomycin in NICUs with its attendant risk for emergence of vancomycin-resistant organisms.
The use of a penicillinase-resistant penicillin antibiotic such as nafcillin to treat a possible staphylococcal infection in this infant is based on the goal of reducing vancomycin use in NICUs. Clinical experience and intervention trials suggest that such a practice is safe. Bloodstream infections caused by CoNS are rarely fulminant or fatal, and they are not associated with an increased case-fatality rate over that seen among uninfected VLBW infants. The clinical outcome of CoNS bacteremia is similar whether the initial antibiotic therapy is vancomycin or another agent that does not reliably treat CoNS infections. In addition, only one of five evaluations for sepsis yields a causative organism. The observation that more than 80% of blood cultures that yield CoNS are positive by 24 hours of incubation makes it possible for the clinician to change antibiotic therapy in a timely fashion if needed. An additional concern of vancomycin therapy has been the association of prior vancomycin use with subsequent development of gram-negative bacteremia among hospitalized pediatric patients. The emergence of community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) in NICUs may limit the use of such a policy in NICUs where the prevalence of CA-MRSA is high. However, by routine screening for MRSA and appropriate isolation precautions for colonized infants, MRSA can be controlled if not eradicated in NICUs.
Aminoglycosides have been the time-honored choice for empiric treatment of infections caused by gram-negative bacilli. Once-daily or extended dosing of gentamicin is used frequently in both full-term and preterm infants based on sound pharmacodynamic and pharmacokinetic considerations. Such a dosing schedule may maximize the bactericidal activity of the aminoglycoside while minimizing its potential toxicity. A retrospective review by Jackson et al. reported the occurrence of hypocalcemia in 3.5% of term and near-term newborns who received gentamicin once daily for 4 days or longer after a change in dosing regimen from every 12 hours to every 24 hours. Although it is known that aminoglycosides enhance urinary calcium excretion, it is not known whether this is potentiated by higher doses of gentamicin.
Aminoglycosides have the distinct advantage of exerting less selective pressure for development of resistance in closed units like the NICU, thus minimizing the risk of emergence of resistant bacteria. This is in contrast to the rapid emergence of cephalosporin resistance when these agents are provided routinely for possible late-onset sepsis. When used for empirical therapy of early-onset infection, cefotaxime has been associated with increased neonatal mortality. However, because CSF penetration of aminoglycosides is poor, their use in meningitis is problematic. If a lumbar puncture is not performed as part of the initial evaluation for possible sepsis, and only an aminoglycoside is used, then effective therapy for gram-negative meningitis may not be provided. Delay in the determination of whether a neonate has meningitis will delay optimal therapy for this condition.
Within 24 hours of collection, the blood cultures yielded gram-negative rods. Cefotaxime was added to the antibiotic regimen. E. coli was subsequently identified from the blood cultures. A lumbar puncture was then performed that demonstrated 4160 white blood cells/mm (90% polymorphonuclear cells, 10% mononuclear cells); 8320 red blood cells/mm 3 ; protein of 433 mg/dL; and glucose of 84 mg/dL (serum glucose of 180 mg/dL). Culture of CSF yielded E. coli.
Question 5: What Is the Treatment of Meningitis in Neonates, Particularly That Caused by Gram-Negative Bacilli?
Table 11.2 provides the recommended antimicrobial treatment for neonatal meningitis based on causative organism. The treatment of gram-negative meningitis initially includes the addition of a third- or fourth-generation cephalosporin such as cefotaxime or cefepime, or a carbapenem antibiotic such as meropenem. Meningitis caused by gram-negative enteric bacilli is challenging because eradication of the organism from CSF is often delayed. Moreover, many of these pathogens are now resistant to ampicillin, and aminoglycoside concentrations are typically low in CSF. Cefotaxime has superior in vitro and CSF bactericidal activity and is the agent of choice. Recently there has been a shortage of cefotaxime in the United States; ceftazidime or cefepime are suitable alternative agents. The cephalosporin agent is combined with an aminoglycoside at least until sterilization of CSF has been achieved. There is no experience or studies using once-daily dosing of aminoglycosides for neonatal meningitis, although from a pharmacodynamic standpoint, such a dosing schedule may be preferred as it should achieve higher CSF concentrations. Continued treatment of gram-negative bacillary meningitis is based on in vitro susceptibility tests. Ampicillin may be used in the infrequent cases when the organism is susceptible.
| Meningitis | Therapy a | Comment |
|---|---|---|
| Initial therapy, CSF abnormal but organism unknown | Ampicillin IV and gentamicin IV, IM and cefotaxime IV | Cefotaxime is added if meningitis suspected or cannot be excluded Alternatives to ampicillin in nursery-acquired infections: vancomycin or nafcillin Alternatives to cefotaxime: ceftazidime, cefepime |
| Bacteroides fragilis spp. Fragilis b | Metronidazole IV | Alternative: meropenem |
| Coliform bacteria c | Cefotaxime IV, IM, and gentamicin | Discontinue gentamicin when clinical and microbiologic response documented Alternative: ampicillin if organism susceptible; meropenem or cefepime for multiresistant organisms Lumbar intrathecal or intraventricular gentamicin usually not beneficial |
| Chryseobacterium ( Flavobacterium ) meningosepticum | Vancomycin IV and rifampin IV, PO | Alternatives: clindamycin, ciprofloxacin |
| Group A streptococcus d | Penicillin G or ampicillin IV | |
| Group B streptococcus b | Ampicillin or penicillin G IV and gentamicin IV, IM | Discontinue gentamicin when clinical and microbiologic response documented |
| Enterococcal spp. d | Ampicillin IV, IM, and gentamicin IV, IM; for ampicillin-resistant organisms: vancomycin and gentamicin | Gentamicin only if synergy documented |
| Other streptococcal species d | Penicillin or ampicillin IV, IM | |
| Gonococcal e | Ceftriaxone IV, IM o r cefotaxime IV, IM | Duration of therapy uncertain (5–10 days?) |
| Haemophilus influenzae d | Cefotaxime IV, IM | Ampicillin if β-lactamase negative |
| Listeria monocytogenes d | Ampicillin IV, IM, and gentamicin IV, IM | Gentamicin is synergistic in vitro with ampicillin but can be discontinued when sterilization achieved |
| Staphylococcus epidermidis (or any coagulase-negative staphylococci) e | Vancomycin IV | Add rifampin if cultures persistently positive Alternative: linezolid |
| Staphylococcus aureus c | MSSA: nafcillin IV MRSA: vancomycin IV | Gentamicin may provide synergy; rifampin if cultures persistently positive |
| Pseudomonas aeruginosa c | Ceftazidime IV, IM and aminoglycoside IV, IM | Meropenem or cefepime are suitable alternatives |
| Candida spp. f | Amphotericin B deoxycholate (AmB-D) X 3–6 weeks | Alternatives: AmB-lipid complex, AmB-liposomal, fluconazole for susceptible strains ( Candida krusei usually resistant) Addition of fluconazole to amphotericin if cultures persistently positive |
| Ureaplasma spp. d | Doxycycline IV or azithromycin IV | Alternatives: ciprofloxacin |
| Mycoplasma hominis d | Clindamycin or doxycycline IV | Alternatives: ciprofloxacin |
Related posts:
Cerebral Circulation and Hypotension in the Premature Infant: Diagnosis and Treatment
Hypothermia for Neonatal Hypoxic-Ischemic Encephalopathy: Different Cooling Regimens and Infants Not Included in Prior Trials
General Supportive Management of the Term Infant With Neonatal Encephalopathy Following Intrapartum Hypoxia-Ischemia
Neonatal-Onset Epilepsies: Early Diagnosis and Targeted Treatment
Hyperbilirubinemia and the Risk for Brain Injury
Amplitude-Integrated EEG and Its Potential Role in Augmenting Management Within the NICU
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
